=================
* Data Analysis *
=================

We used three different statistical packages (i.e., SPSS v.12, Microsoft Excel 2003, and Stata v.9) for our analysis on the datasets. For the descriptive statistics including frequencies and transition probabilities (e.g., Tables 1-9 and Figures 1-6), both SPSS v.12 and Microsoft Excel 2003 were used.  We also used Stata v.9 for the Logit regression analysis.


====================
* File Description *
====================

1. Dominicks_AllStores.sav  (SPSS file): Dominicks dataset of all stores for the following 29 product categories.

PRODCAT	Description
===========================
1	Analgesics
2	Bath Soap
3	Bathroom Tissue
4	Beer
5	Bottled Juice
6	Canned Soup
7	Canned Tuna
8	Cereals
9	Cheeses
10	Cigarettes
11	Cookies
12	Crackers
13	Dish Detergents
14	Fabric Softeners
15	Front-End-Candies
16	Frozen Dinners
17	Frozen Entrees
18	Frozen Juices
19	Grooming Products
20	Laundry Detergents
21	Oatmeal
22	Paper Towels
23	Refrigerated Juices
24	Shampoos
25	Snack Crackers
26	Soaps
27	Soft Drinks
28	Toothbrushes
29	Toothpastes
===============================

2. Dominicks_4Stores.sav  (SPSS file): Dominicks dataset of 4 stores (i.e., Store 8, 12, 122, and 133).

3. Dominicks_4Stores_PCH_MK.sav (SPSS file): Dominicks dataset of price changes (i.e., PCH = 1) for 4 stores with Markov chain variables (i.e., MKEND1, MKEND2)

/* Markov Chain Variables (MKEND1, MKEND2) */

STRING strprvprice (A8).
  COMPUTE strprvprice = STRING(prvprice, F5.2).
  EXECUTE.

strprvprice = RTRIM(LTRIM(strprvprice))
prvlength = length(RTRIM(LTRIM(strprvprice)))
strprvend1 = SUBSTR(strprvprice,prvlength)
strprvend2 = SUBSTR(strprvprice,prvlength-1)

prvend1 = NUMBER(strprvend1,F1.0)
prvend2 = NUMBER(strprvend2,F2.0)

mkend1 = prvend1 * 10 + end1
mkend2 = prvend2 * 100 + end2

/* Price Ending of a Price Change */

STRING strpch (A8).
  COMPUTE strpch = STRING(abspch, F5.2).
  EXECUTE.

strpch = RTRIM(LTRIM(strpch))
pchlength = length(RTRIM(LTRIM(strpch)))
strpchend1 = SUBSTR(strpch,pchlength)
strpchend2 = SUBSTR(strpch,pchlength-1)
if strpchdend2 = null then strpchdend2 = strpchdend1

pchend1 = NUMBER(strpchend1,F1.0)
pchend2 = NUMBER(strpchend2,F2.0)
mdimes = 0
mdimes = 1 if pchend1 = 0 // Price Changes in Multiple of Dimes
mdollars = 0
mdollars= 1 if pchend2 = 0 // Price Changes in Multiple of Dollars

4. Dominicks_4Stores_Regular_Price_Logit.txt  (for Stata): Dominicks dataset of 4 stores with regular price for Logit regression

5. Dominicks_HighQ.sav and Domminicks_LowQ.sav  (SPSS file): Dominicks dataset for the low and high quartile of products in terms of sales volume

6. Dominicks_All_PCH_MK.sav (SPSS file): Dominicks dataset of price changes (i.e., PCH = 1) for all stores with Markov chain variables (i.e., MKEND1, MKEND2)

7. Internet_Data.sav  (SPSS file): Internet datasets for 293 retailers with the following 10 product categories

PRODCAT	Description
=======================
1	Music CDs
2	Movie DVDs
3	Video Games
4	Software
5	Hard Drives
6	PDAs
7	DVD Players
8	PC Monitors
9	Digital Cameras
10	Notebook PCs
========================

/* Price Ending of a Price */

STRING strprice (A8).
  COMPUTE strprice = STRING(price, F7.2).
  EXECUTE.

strprice = RTRIM(LTRIM(strprice))
plength = length(RTRIM(LTRIM(strprice)))
strend1 = SUBSTR(strprice,plength)
strend2 = SUBSTR(strprice,plength-1)
strend3 = SUBSTR(strprice,plength-3)
strend4 = SUBSTR(strprice,plength-4)
strdend1 = SUBSTR(strprice,plength-3, 1)
strdend2 = SUBSTR(strprice,plength-4, 2)

end1 = NUMBER(strend1,F1.0)
end2 = NUMBER(strend2,F2.0)
end3 = NUMBER(strend3,F4.0)
end4 = NUMBER(strend4,F5.0)
dend1 = NUMBER(strdend1,F1.0)
dend2 = NUMBER(strdend2,F2.0)

8. Internet_Data_PCH_MK.sav  (SPSS file): Internet datasets of price changes (i.e., PCH = 1) with Markov chain variables (i.e., MKEND1, MKEND2, MKDEND1, MKDEND2)

/* Markov Chain Variables (mkend1, mkend2, mkdend1, mkdend2) */

STRING strprvprice (A8).
  COMPUTE strprvprice = STRING(prvprice, F7.2).
  EXECUTE.

strprvprice = RTRIM(LTRIM(strprvprice))
prvlength = length(RTRIM(LTRIM(strprvprice)))
strprvend1 = SUBSTR(strprvprice,prvlength)
strprvend2 = SUBSTR(strprvprice,prvlength-1)
strprvdend1 = SUBSTR(strprvprice,prvlength-3, 1)
strprvdend2 = SUBSTR(strprvprice,prvlength-4, 2)

prvend1 = NUMBER(strprvend1,F1.0)
prvend2 = NUMBER(strprvend2,F2.0)
prvdend1 = NUMBER(strprvdend1,F1.0)
prvdend2 = NUMBER(strprvdend2,F2.0)

mkend1 = prvend1 * 10 + end1
mkend2 = prvend2 * 100 + end2
mkdend1 = prvdend1 * 10 + dend1
mkdend2 = prvdend2 * 100 + dend2

/* Price Ending of a Price Change */

STRING strpch (A8).
  COMPUTE strpch = STRING(abspch, F7.2).
  EXECUTE.

strpch = RTRIM(LTRIM(strpch))
pchlength = length(RTRIM(LTRIM(strpch)))
strpchend1 = SUBSTR(strpch,pchlength)
strpchend2 = SUBSTR(strpch,pchlength-1)
strpchdend1 = SUBSTR(strpch,pchlength-3, 1)
strpchdend2 = SUBSTR(strpch,pchlength-4, 2)

pchend1 = NUMBER(strpchend1,F1.0)
pchend2 = NUMBER(strpchend2,F2.0)
pchdend = NUMBER(strpchdend1,F1.0)
pchdend2 = NUMBER(strpchdend2,F2.0)

mdimes = 0
mdimes = 1 if pchend1 = 0 // Price Changes in Multiple of Dimes
mdollars = 0
mdollars= 1 if pchend2 = 0 // Price Changes in Multiple of Dollars
m10 = 0
m10 = 1 if pchend2 = 0  & pchdend1 = 0 // Price Changes in Multiple of $10
m100 = 0
m100 = 1 if pchend2 = 0  & pchdend2 = 0 // Price Changes in Multiple of $100

9. Internet_Data_Logit.txt (for Stata): Internet dataset for Logit regression


============================
* Description of Variables *
============================

Variable	Description
===========================================================
ID		ID generated from the combination of STORE and PID
PRODCAT		Product category
STORE		Store ID
PID		Product ID converted from UPC code
WEEK		Week of sales
DAY		Day of sales
PRICE		Transaction price
END1		The last digit of PRICE
END9		Dummy variable indicating whether END1 is "9" or not
END2		The last two digits of PRICE
END99		Dummy variable indicating whether END2 is "99" or not
END3		The last three digits of PRICE
END999		Dummy variable indicating whether END3 is "9.99" or not
END4		The last four digits of PRICE
END9999		Dummy variable indicating whether END is "99.99" or not
DEND1		The last dollar digit of PRICE
DEND9		Dummy variable indicating whether DEND1 is "9" or not
DEND2		The last two dollar digits of PRICE
DEND99		Dummy variable indicating whether DEND2 is "99" or not
SALE		Dummy variable indicating whether a product is on sale or not
PCHANGE		The amount of price change from the previous week (or day)
ABSPCH		The absolute value of PCHANGE
PCH		Dummy variable indicating whether a price changed from the previous week (or day) or not
PCHEND1		The last digit of PCHANGE
MDIMES		Dummy variable indicating whether a price change is in multiple of dimes (i.e., PCHEND1=0) 
PCHEND2		The last two digits of PCHANGE
MDOLLARS	Dummy variable indicating whether a price change is in multiple of dollars (i.e., PCHEND2=0)
PCHDEND1	The last dollar digit of PCHANGE
M10		Dummy variable indicating whether a price change is in multiple of $10 (i.e., PCHDEND1=0) 
PCHDEND2	The last two dollar digits of PCHANGE
M100		Dummy variable indicating whether a price change is in multiple of $100 (i.e., PCHDEND2=0)
PRVPRICE	A price of previous week (or day)
PRVEND1		The last digit of PRVPRICE
PRVEND2		The last two digits of PRVPRICE
PRVDEND1	The last dollar digit of PRVPRICE
PRVDEND2	The last two dollar digits of PRVPRICE
MKEND1		The combination of PRVEND1 and END1 (i.e., PRVEND1 * 10 + END1)
MKEND2		The combination of PRVEND2 and END2 (i.e., PRVEND2 * 100 + END2)
MKDEND1		The combination of PRVDEND1 and DEND1 (i.e., PRVDEND1 * 10 + DEND1)
MKDEND2		The combination of PRVDEND2 and DEND2 (i.e., PRVDEND2 * 100 + DEND2)
LOWCAT		Dummy variable indicating whether a product is in low price categories in Internet dataset, 		such as Music CDs, Movie DVDs, and Video Games (i.e., PRODCAT = 1, 2, or 3) 
LOWQ		Dummy variable indicating whether a product is in the low quartile in terms of ��9��-ending 		popularity or not
======================================================================





